A Distributed Electric Vehicles Charging System Powered by Photovoltaic Solar Energy with Enhanced Voltage and Frequency Control in Isolated Microgrids

Abstract

This study presents a photovoltaic (PV)-based electric vehicle (EV) charging system designed to optimize energy use and support isolated microgrid operations. The system integrates PV panels, DC/AC, AC/DC, and DC/DC converters, voltage and frequency droop control, and two energy management algorithms: Power Sharing and SEWP (Spread Energy with Priority). The DC/AC converter demonstrated high efficiency, with stable AC output and Total Harmonic Distortion (THD) limited to 1%. The MPPT algorithm ensured optimal energy extraction under both gradual and abrupt irradiance variations. The DC/DC converter operated in constant current mode followed by constant voltage regulation, enabling stable power delivery and preserving battery integrity. The Power Sharing algorithm, which distributes PV energy equally, favored vehicles with a higher initial state of charge (SOC), while leaving low-SOC vehicles at modest levels, reducing satisfaction under limited irradiance. In contrast, SEWP prioritized low-SOC EVs, enabling them to achieve higher SOC values compared to the Power Sharing algorithm, reducing SOC dispersion and enhancing fairness. The integration of voltage and frequency droop controls allowed the station to support microgrid stability by limiting reactive power injection to 30% of apparent power and adjusting charging current in response to frequency deviation.

1. Introduction

EVs are a promising solution to mitigate the environmental impacts caused by the transport sector. The replacement of internal combustion vehicles (ICV) with EVs can significantly reduce the emission of atmospheric pollutants such as carbon dioxide (CO₂), nitrogen oxides (NO), and particulate matter (PM), which are responsible for public health issues such as respiratory and cardiovascular diseases, as well as contributing to global warming [1,2,3]. The absence of exhaust emissions in EVs reduces the concentration of harmful pollutants in urban areas, improving air quality and public health. Additionally, EVs are more efficient in terms of energy conversion, utilizing a higher proportion of stored energy to move the vehicles compared to combustion engines [3,4,5,6,7].

The proliferation of EVs marks a significant stride towards achieving a cleaner and more sustainable energy matrix [8]. However, this growth poses challenges regarding the electrical grid’s capacity to meet the rising energy demand for recharging these vehicles, particularly during peak periods. Vehicle-to-grid (V2G) technology offers a vital solution for energy availability within microgrids, enabling EVs to stabilize the grid by returning energy, thus functioning as controlled loads and mobile storage units [9]. Furthermore, decentralized energy resources (DER) and sustainable charging stations address grid capacity concerns by enhancing flexibility through localized generation and storage. This approach ensures that the energy is produced and consumed closer to its point of use, alleviating the pressure on transmission and distribution grids [8,10,11,12,13,14].

The massive introduction of EVs and renewable energy sources changes the energy consumption pattern, making it less predictable and potentially leading to failures in the energy production system. The DERs are an alternative solution to the traditional concept of centralized production, offering new opportunities to improve the current electrical energy system. It allows mitigation of the transmission grid, creates conditions for economic development, makes the consumer more aware of energy wastage, and promotes the production of cleaner energy. It is necessary to continue investing in the development of DERs to ensure greater reliability and quality of service, reducing grid congestion and increasing energy performance [15].

To maximize the benefits of EVs in reducing atmospheric pollution, it is essential to establish clear and achievable goals [4,5,8]. Some of the recommended targets include increasing the penetration of EVs through subsidies, tax incentives, and policies supporting charging infrastructure, decarbonizing energy generation to ensure that the electricity used to charge EVs comes from renewable sources, and developing fast and smart charging infrastructures to support the growing demand for EVs and enable integration with distributed generation. The promotion of smart charging technologies and the implementation of optimized charging strategies are also fundamental to reducing peak demand on the electrical grid [16,17,18].

The primary challenge in charging EVs with renewable sources, such as solar energy, is location. For solar energy to be used at its maximum on buildings and houses, the EVs need to be at home during peak solar production hours; however, most people are at work during this time, making it difficult to use this energy [19]. Additionally, the lack of parking spaces in some buildings and homes presents a significant obstacle for EV charging, potentially leading to infractions as individuals seek ways to charge their vehicles.

Daytime EV charging can be based on photovoltaic solar energy, given its production profile, assuming that charging during this period often takes place at workplaces or car parks during working hours [20]. That said, the solution must pass through a PV-powered charging station (PVCS), designed for charging EVs, using PV panels installed on car parking shades and building rooftops [21]. A paper proposed an optimal charging strategy in order to verify the economic and environmental impacts of a PVCS [22]. The method determines the total charging power of EVs in the parking garage to optimize the power drawn from the grid. The charging strategies of multiple Plug-in Hybrid Electric Vehicles (PHEVs) in PV-equipped apartment buildings and office buildings are discussed in [23,24]. PV energy allows for the reduction in production needs from the rest of the electrical system and meets the charging needs of EVs when their number is not high [5,25]. The integration of a stationary storage battery is essential to ensure the reliable operation of the PVCS, particularly when the power demand from EVs exceeds the generation capacity of the PV panels or during periods of low or absent solar irradiance [26,27,28,29]. A paper describes a PV-based charging station equipped with a battery energy storage system. The test results validate the enhanced power quality operation of the EV charging station under dynamic conditions, including intermittent PV isolation and several modes of battery energy storage. Also examined in depth is the system’s response during battery charging and discharging [30]. Current studies indicate that investing in batteries is economically unfeasible due to their high costs; however, forecasts suggest that in the future, battery costs will decrease significantly, making them a viable solution for energy storage and management [27].

Although PVCS theoretically requires a battery, it is possible to charge a high number of EVs without consuming much energy from the grid or resorting to the energy provided by a battery (minimizing the cost of the PVCS) and still ensuring that all EVs receive enough energy. For that to become possible, it is imperative to manage the energy produced by the PV and distribute it efficiently among the different EVs. That management can only be performed by an algorithm implemented on the control center (as shown in Figure 1) of the PVCS that can control and monitor the energy flows [31,32,33]. A real-time energy management algorithm for a grid-connected charging park in an industrial or commercial workplace is developed in [34]. Statistical and forecasting models are used to model various uncertainties, including the PV power, PHEVs’ arrival and departure time, and the available energy of the batteries. To address the limitations in large-scale EV charging, heuristic algorithms inspired by natural processes, such as Particle Swarm Optimization (PSO) [35], Chemical Reaction Optimization (CRO) [36], and Genetic Algorithm (GA) [37], have been employed. These artificial intelligence algorithms have demonstrated the potential to outperform conventional methods in terms of computational overhead, yielding promising results on charging EVs through the grid [38].

A feasibility study of a PVCS has been conducted in [32] by analyzing its effectiveness based on technical, economic, and environmental aspects by comparing the impact of different geographical areas on the installation location. The study investigates how a PVCS can contribute to charging EVs with different energy mixes and compares the produced CO₂ emissions of charging EV batteries solely from the grid, from the PVCS, and with internal combustion engine vehicles. They have found that the PVCS concept is more efficient in countries with high annual average irradiance and significant CO₂ emissions in their grid, but it remains economically unfeasible due to expensive storage systems [33].

A user-friendly smart charging method, which includes interactions with EV users via an interface, has been developed in [39], where the EV user is a key player in the process of choosing the best scenario among uncoordinated charging, smart charging, and bidirectional smart charging control in a PVCS. The proposed methodology is based on real-time rule-based control and predictive linear optimization control. The results showed that bidirectional charging control had the best cost reduction, while uncoordinated charging control costs the most.

Several studies have presented control mechanisms aimed at minimizing the discomfort of EV users at charging stations equipped with PV sources and connected to a public grid, without a stationary storage system. However, our approach distinguishes itself in several ways. Firstly, while some studies allow only one charging mode, such as the slow mode, our research introduces three different charging modes. Additionally, while the focus of other works is on maximizing social welfare and minimizing EV user discomfort, our objective is to maximize the use of energy produced by PV for EV charging, thereby reducing grid energy consumption. Previous studies have focused on solar charging stations with batteries, but our work demonstrates that EV charging can be effectively managed without batteries, as the charging period coincides with photovoltaic production hours. Furthermore, energy management studies for EV charging have primarily addressed grid energy management; however, those algorithms can also be applied to manage energy produced by PV systems, providing an efficient management of the power produced by the PV panels. Therefore, the main contributions of this paper are as follows:

(1)

In order to evaluate the long-term effectiveness of the commercialized PVCS, a set of evaluation indexes are introduced, including the quality of service, economic benefits, environmental benefits, and impacts on the power grid.

(2)

The developed project aims to create an innovative solar charging park that eliminates the need for storage batteries. By utilizing advanced solar energy conversion and distribution technology, the park ensures that all electric vehicles can be charged efficiently and sustainably. This system represents a significant advancement in electric vehicle charging infrastructure, promoting the adoption of renewable energy and contributing to the reduction of carbon emissions.

(3)

The power distribution algorithms of the PV panels developed in the project allow for the maximum utilization of available solar energy while ensuring customer satisfaction by considering the desired SOC. With electric vehicles parked, it is possible to establish distinct charging priorities based on the duration of the vehicles’ parking time and the users’ needs. This system offers a low cost for users while providing a higher return for energy producers.

(4)

The solar charging station developed in the project enhances the sustainability of buildings by eliminating the impact of electric vehicle charging on the electrical grid, as consumption from the grid is greatly reduced or even eliminated. DER production can cause frequency and voltage fluctuations in microgrids. However, in this solution, the PV system has minimal impact on the microgrid and can even correct these fluctuations, ensuring more stable and efficient operation.

This paper is structured as follows: Section 2 presents the components of the PVCS, more precisely, the DC/AC converter that connects the PV to the grid, and the components of the EV charger connected to the grid. Section 3 presents three different charging options, more precisely, the possibility of charging with energy exclusively from the photovoltaic panel or the electrical power grid, or a combination of both. It also presents two different algorithms used to perform intelligent energy management and distribution in order to maximize the use of solar energy and minimize dependence on the grid, which in turn leads to a relief of the microgrid overload. Simulation results to demonstrate and verify the effectiveness of the algorithms developed are described in Section 4. Finally, the conclusions are given in Section 5.

2. Proposed System

The PVCS we propose is shown in Figure 2. It consists of a grid converter, PV panels, and an EV charger. The grid-connected converter permits power exchange between the DC and AC sides. The DC/AC converter integrates Perturb and Observe (P&O) Maximum Power Point Tracking (MPPT) capabilities to maximize power extraction from the PV panels [40]. The AC/DC converter and the DC/DC converter secure a steady power supply to the EV batteries, depending on the system operating parameters (they specifically depend on the algorithms used in the control center).

A unitary power factor is assumed solely for the purpose of system sizing. This assumption simplifies the design process without influencing the control strategy or the dynamic performance of the system.

2.1. DC/AC Converter

The DC/AC converter, which interfaces the PV panels with the isolated AC microgrid, employs a multi-layered control architecture to ensure stable and efficient operation. This architecture includes current control, voltage control, MPPT control, and a voltage droop control mechanism.

The current control ensures that the output AC currents ($i_{o1},i_{o2},i_{o3}$) are in phase with the microgrid voltages ($v_{o1},v_{o2},v_{o3}$), and follow the reference currents ($i_{o1ref},i_{o2ref},i_{o3ref}$) generated by the voltage control loop. The voltage control regulates the DC-link voltage ($V_{DC}$) at the PV terminals, enabling the extraction of maximum power from the PV panels by tracking the DC reference voltage provided by the MPPT, while simultaneously ensuring that the DC-link voltage remains higher than the AC-side voltage to preserve converter stability. The MPPT control continuously adjusts the operating point of the PV panels to maximize power extraction, based on real-time measurements of the PV voltage ($V_{DC}$) and current ($i_{pv}$). The AC voltage droop control contributes to the regulation of the isolated microgrid voltage by adjusting the reactive power injected by the converter. This control generates a phase angle ($\varphi$) for each output current, which is subsequently processed by a synchronizer to ensure proper alignment with the grid voltage [41].

The dynamics of the DC/AC converter can be written considering that each k-leg (k ∈ {1, 2, 3}) has two IGBT switches responsible for converting the DC current into an AC current. Assuming that when switch S_k1 is closed, then S_k1 = 1; otherwise, switch S_k2 is open, S_k2 = 0, while S_k2 =1 − S_k1. Each k-leg outputs two voltages (states), defined by a variable γ_k, dependent on the leg switches states.

$${\gamma }_k=\{\begin{array}{c}1\to S_{k1}=1 \mathrm{and} S_{k2}=0\\ 0\to S_{k1}=0 \mathrm{and} S_{k2}=1\end{array}$$

(1)

2.1.1. Grid Current Controller

To control the current injected into the electrical isolated microgrid, it is necessary to extract the sinusoidal signal of each phase to ensure that the injected current is in phase with the grid voltage. This process is carried out using phase-locked loops (PLLs).

The extracted phase is then multiplied by the RMS current, $i_{oRMS}$, produced by the DC voltage controller, and this operation generates the current $i_{o1ref}$ for phase 1, given by Equation (2), which is subsequently compared with the AC current. Finally, the result of this comparison is fed into a current controller, which is responsible for generating the command signal. This signal, when compared with a triangular carrier, is applied to the IGBTs to control the current injected into the grid, as illustrated in Figure 3.

$${ i}_{o1ref}=\sqrt{2}{ i}_{oRMS}sin\left(\omega t-\varphi \right)$$

(2)

Therefore, the currents $i_{o1}$, $i_{o2}$ and $i_{o3}$ can be regulated by adjusting the inverter output voltage $V_k$ [42]. By applying Kirchhoff’s voltage law to the three-phase circuits, the following expressions are obtained:

$$\begin{gathered} {\displaystyle \frac{d{i}_{o1}}{dt}}=-{\displaystyle \frac{R_L}{L}}{i}_{o1}-{\displaystyle \frac{v_{o1}}{L}}+{\displaystyle \frac{V_1}{L}} \\ {\displaystyle \frac{d{i}_{o2}}{dt}}=-{\displaystyle \frac{R_L}{L}}{i}_{o2}-{\displaystyle \frac{v_{o2}}{L}}+{\displaystyle \frac{V_2}{L}} \\ {\displaystyle \frac{d{i}_{o3}}{dt}}=-{\displaystyle \frac{R_L}{L}}{i}_{o3}-{\displaystyle \frac{v_{o3}}{L}}+{\displaystyle \frac{V_3}{L}} \end{gathered}$$

(3)

The inverter output voltages $V_1$, $V_2$, and $V_3$ are determined by the leg switches and the inverter input voltage $V_{DC}$ [42]. By applying the Laplace Transform (LT) to Equation (3) with this modification, we obtained:

$$\begin{gathered} i_{o1}={\displaystyle \frac{\frac{1}{L}}{s+\frac{R_L}{L}}}\left(-v_{o1}+{\displaystyle \frac{2{\gamma }_1-{\gamma }_2-{\gamma }_3}{3}}{V}_{DC}\right) \\ {i}_{o2}={\displaystyle \frac{\frac{1}{L}}{s+\frac{R_L}{L}}}\left(-v_{o2}+{\displaystyle \frac{2{\gamma }_2-{\gamma }_1-{\gamma }_3}{3}}{V}_{DC}\right) \\ {i}_{o3}={\displaystyle \frac{\frac{1}{L}}{s+\frac{R_L}{L}}}\left(-v_{o3}+{\displaystyle \frac{2{\gamma }_3-{\gamma }_1-{\gamma }_2}{3}}{V}_{DC}\right) \end{gathered}$$

(4)

Using Equation (4), the control gains of the PI controller responsible for controlling the current of phase 1 can be computed based on the block diagram shown in Figure 4. For the other two phases, the block diagram is the same as Figure 4, with the difference in the leg switches indexed to each phase.

The proportional integral controller (PI) is given by

$$PI=K_{p i_{o1}}+{\displaystyle \frac{K_{i i_{o1}}}{s}}$$

(5)

To determine the values of $K_{p i_{o1}}$ and $K_{i i_{o1}}$, it is necessary to analyze the closed-loop feedback system (Figure 4) and establish the relation between the phase 1 current, $i_{o1}$, and the reference current, $i_{o1ref}$, thereby obtaining the expression for $i_{o1}$ shown below:

$$\begin{gathered} i_{o1} ={\displaystyle \frac{\left( -v_{o1}+\frac{-{\gamma }_2-{\gamma }_3}{3}{V}_{DC} \right)\frac{1}{L}s}{s^{2 }+ \left(\frac{R_L}{L} + \frac{2 V_{DC}}{3L} K_{P i_{o1}}\right)s + \frac{2 V_{DC}}{3 L} K_{i i_{o1}}}} \\ +{\displaystyle \frac{\frac{2 V_{DC}}{3 L} K_{P i_{o1}}\left(s + \frac{ K_{i i_{o1}}}{ K_{P i_{o1}}}\right)}{s^{2 }+ \left(\frac{R_L}{L} + \frac{2 V_{DC}}{3L} K_{P i_{o1}}\right)s + \frac{2 V_{DC}}{3 L} K_{i i_{o1}}}} i_{o1ref} \end{gathered}$$

(6)

By applying the limit as $s$ tends to zero, it can be observed that the current $i_{o1}$ depends on the reference current $i_{o1ref}$, confirming that under steady-state conditions, the current follows the reference. Comparing the denominator of the closed-loop transfer function (6) with the canonical form of a second-order system [43,44], the following equations for $K_{p i_{o1}}$ and $K_{i i_{o1}}$ are obtained:

$$\{\begin{array}{c} 2\xi {\omega }_n = {\displaystyle \frac{R_L}{L}} + {\displaystyle \frac{2{ V}_{DC} K_{P i_{o1}}}{3L}} \iff K_{P i_{o1}} = {\displaystyle \frac{3L}{2V_{DC}}} \left(2\xi {\omega }_n - {\displaystyle \frac{R_L}{ L}}\right)\\ {\omega }_n^2 = {\displaystyle \frac{2 V_{DC}{ K}_{i i_{o1}}}{3L}} \iff K_{i i_{o1}} = {\displaystyle \frac{3L}{2V_{DC}}}{\omega }_n^2\end{array}$$

(7)

Thus, the expressions presented in (7), where ${\omega }_n$ represents the natural frequency of the second-order system and the variable $\xi$ represents the damping factor. The lower the damping factor, the faster the current stabilizes, with the drawback that if the damping factor is less than ${\displaystyle \frac{1}{\sqrt{2}}}$, the system exhibits overshoots before stabilizing, potentially reaching very high maximum values. Therefore, it is advisable to use values around ${\displaystyle \frac{1}{\sqrt{2}}}$, not too high to avoid prolonged stabilization times, not too low to prevent overshoots [44].

2.1.2. DC Voltage Controller

The grid converter is responsible for converting the power produced by the PV module into the EV charger. For the grid converter to work correctly, the voltage at capacitor C should follow the reference provided by the MPPT algorithm and be maintained at a level higher than the maximum RMS grid voltage ($v_{ok}$), with $k$ being the index of each grid phase, to be able to impose the direction of power flow through inductor L.

$$V_{DCref}>\sqrt{2}{v}_{ok}$$

(8)

To guarantee that the voltage at capacitor C is maintained at $V_{DCref}$, while simultaneously injecting power into the EV charger, a PI controller should be devised.

Applying Kirchhoff’s current law to the capacitor of the DC-link, we obtain:

$${\displaystyle \frac{d{V}_{DC}}{dt}}={\displaystyle \frac{ 1}{C}} \left(i_{pv} - i_d\right)$$

(9)

where $i_d$ is the current that flows into the DC/AC converter and $i_{pv}$ is the current produced by the PV panels. The power produced by the PV panels is approximately equal to the output power of the converter. That being said, $i_d$ current can be given by ${\displaystyle \frac{\eta 3 v_{oRMS} i_{oRMS}}{V_{DCref}}}$, $i_{oRMS}$ being the effective value of the AC current of the grid converter, $v_{oRMS}$ being the effective value of the AC voltage of the grid, and $\eta$ being the efficiency of the system. Applying the expression to the current $i_d$ and the Laplace transform to Equation (9), the following result is obtained:

$$V_{DC}={\displaystyle \frac{1}{sC}}\left(i_{pv}-{\displaystyle \frac{\eta 3 v_{oRMS} i_{oRMS}}{V_{DCref}}}\right)$$

(10)

By using Equation (10), the control gains of the PI controller, which is responsible for regulating the voltage of the DC-link capacitor, can be determined from a block diagram similar to Figure 4. The closed-loop function transfer function (TF) of the system is given by

$$\begin{gathered} V_{DC}={\displaystyle \frac{ \frac{1}{C}\left(i_{pv}\right)s}{s^2-\left(\frac{\eta 3 v_{oRMS}}{V_{DCref}} \right)\frac{K_{P V_{DC}}}{C}s-\frac{\eta 3 v_{oRMS}}{V_{DCref}} \frac{K_{i V_{DC}}}{C}}} \\ -{\displaystyle \frac{ \frac{\eta 3 v_{oRMS}}{V_{DCref}} \frac{K_{P V_{DC}}}{C}\left(s+\frac{K_{i V_{DC}}}{K_{P V_{DC}}}\right)}{s^2-\left(\frac{\eta 3 v_{oRMS}}{V_{DCref}} \right)\frac{K_{P V_{DC}}}{C}s-\frac{\eta 3 v_{oRMS}}{V_{DCref}} \frac{K_{i V_{DC}}}{C}}}{ V}_{DCref} \end{gathered}$$

(11)

If the limit is calculated when s → 0 (steady state), it is seen that the input voltage tends towards the reference value, verifying that $i_{pv}$ has low influence in the steady state. To help determine the values of the PI compensator parameters, it is imperative to determine the relation between the closed-loop equation of the voltage control system ${ V}_{DC}$ (Equation (11)) and the second-order equations in canonical form [43,44]. After comparing Equation (11) and the second-order equations in canonical form, the following equations for $K_{P V_{DC}}$ and $K_{i V_{DC}}$ are obtained:

$$\{\begin{array}{c} K_{P V_{DC}} = - {\displaystyle \frac{V_{DCref}}{\eta 3 v_{oRMS}}}C 2{\xi }_{V_{DC}}{{ \omega }_n}_{V_{DC}}\\ K_{i V_{DC}} = - {\displaystyle \frac{V_{DCref}}{\eta 3 v_{oRMS}}}C{ \omega }_{n{V}_{DC}}^2\end{array}$$

(12)

where ${{\omega }_n}_{V_{DC}}$ (natural frequency) is defined as $2\pi$ divided by the period corresponding to the natural frequency of the second-order system $\left({\displaystyle \frac{2\pi }{T_{V_{DC}}}}\right)$, and the variable ${\xi }_{V_{DC}}$ describes the damping coefficient of the second-order system. As with the current compensator, it is advisable to use damping factor values around ${\displaystyle \frac{1}{\sqrt{2}}}$ to ensure the system responds quickly to variations in reference voltage without overshoots [44].

2.1.3. AC Voltage Droop Controller

A voltage droop control strategy is employed in the inverter of the solar charging station to regulate the reactive power exchange with the electrical isolated microgrid and thereby contributes to local voltage stabilization. This decentralized method enables the inverter to autonomously respond to voltage deviations at the point of common coupling (PCC), without requiring communication with other grid elements [45]. Since the reactive power support is usually limited to approximately 30% of the apparent power, a single station cannot fully control the isolated microgrid voltage. Instead, the mechanism contributes to its regulation, and an effective stabilization requires multiple charging stations operating simultaneously, whose combined reactive power injection or absorption progressively drives the isolated microgrid voltage closer to the reference value [41,46].

In this strategy, the DC/AC converter continuously monitors the RMS grid voltage, $v_{oRMS}$, and compares it to a predefined reference value $v_{oRMSref}$. Based on this deviation, the controller generates a phase angle $\varphi$, which is proportional to the voltage difference and nominal system voltage, $V_N$. This angle is then applied to the converter’s current control loop.

$$\varphi =K_V{\displaystyle \frac{v_{oRMSref}-v_{oRMS}}{V_N}}$$

(13)

The phase angle $\varphi$ defines the reactive power exchanged with the isolated microgrid. By adjusting this angle, the inverter modulates its current injection to either supply or absorb reactive power, thereby contributing to the regulation of the local voltage profile. The voltage regulation gain $K_V$ is obtained by multiplying the maximum phase angle $\varphi$ by the nominal voltage $V_N$, and then dividing the result by the isolated microgrid voltage variation ($v_{oRMSref}-v_{oRMS}$) [46].

2.1.4. MPPT (Maximum Power Point Tracking)

MPPT is a critical technique employed in photovoltaic systems to optimize the extraction of energy from solar panels. MPPT continuously adjusts the electrical DC voltage within the circuit to ensure that the panels operate at their maximum power point, irrespective of fluctuations in irradiance and temperature. This adjustment is achieved through constant monitoring of the voltage and current of the panels, enabling the system to swiftly adapt to environmental changes and maintain peak efficiency. The implementation of MPPT in photovoltaic systems is vital for enhancing energy generation, reducing operational costs, and extending the lifespan of solar panels [40].

A common MPPT technique used in PV systems is the P&O algorithm. It works by slightly perturbing the PV module’s operating point (typically the voltage) and observing the corresponding change in power output. The algorithm then adjusts the operating point based on whether the power increased or decreased with the perturbation, essentially “climbing” the power curve to find the MPP (Maximum Power Point), as shown in Figure 5 by the dashed red lines [40].

2.2. Battery Charging System for EVs

The battery charging system with a DC-DC buck converter is depicted in Figure 6. The converter comprises an IGBT, a diode, and an inductor, $L_{DC}$, where i_L represents the current that flows through the inductor and a capacitor, $C_{DC}$, where $V_{EV}$, denotes the voltage of the capacitor, and it also represents the output voltage of the power converter across the EV battery since the battery is connected in parallel with the capacitor. The control signal γ is a discrete variable that assumes values, either 1 (ON) or 0 (OFF), to command the switching state of the transistor Q. Usually, a pulse with a modulation (PWM) system is used to generate the discrete signal γ.

The equivalent electrical model of the EV battery consists of a controlled voltage source that characterizes the state-of-charge (SOC) behavior as a function of the open-circuit voltage, which exhibits an almost linear performance when the state of charge is between 20% and 80%, but becomes nearly exponential outside this range. For this reason, the battery charging process is performed in current mode until the voltage reaches the nominal value, and then the battery is charged in voltage mode to ensure that in the exponential zone the voltage at the battery terminals is well regulated [47,48].

In order for the DC/DC buck converter to be used for charging an EV battery with variable charging power, it is necessary to introduce a limiter at the output of the PI Voltage Controller (as shown in Figure 6). This limiter has a maximum limit set by the current, $i_{algorithm}$, derived from the charging power (charging power divided by the battery voltage). Thus, the battery is initially charged with the current defined by the available charging power until the voltage reaches its maximum (typically corresponding to around 80% of the SOC). At this point, the voltage controller becomes active and, in order to prevent the battery voltage from rising further and ensure its protection, it progressively reduces the charging current through voltage regulation. From this voltage level, the PI Voltage Controller operates to ensure that the battery voltage does not exceed the maximum value. The system integrates the frequency droop control mechanism, which is responsible for adjusting the power available for EV charging according to the isolated microgrid frequency deviations: when the frequency decreases, part of the active power is diverted to the isolated microgrid, thereby reducing the charging power available; conversely, when the frequency increases, the converter absorbs active power from the isolated microgrid, thus enhancing the power available for charging the EVs [41].

Assuming that the DC/DC buck converter is operating in continuous conduction mode, the dynamic model of the converter can be written as follows:

$${\displaystyle \frac{d{i}_L}{dt}}={\displaystyle \frac{1}{L_{DC}}}\left(\mathsf{\gamma }U-V_{EV}\right)$$

(14)

$${\displaystyle \frac{d{V}_{EV}}{dt}}={\displaystyle \frac{1}{C_{DC}}}\left(i_L-i_{EV}\right)$$

(15)

2.2.1. Inductor Current Controller

So that the inductor current follows the current defined by the battery voltage controller, $i_{Lref}$, a PI controller should be devised. To design that controller, Equation (14) is used, which expresses a relation between the variation in the current i_L and the input and output voltages; applying the Laplace transform, we obtain the following equation:

$$i_L={\displaystyle \frac{1}{s{L}_{DC}}}\left(\mathsf{\gamma }U-V_{EV}\right)$$

(16)

Equation (16) describes the system dynamics in the Laplace domain and serves as the basis for designing the PI current controller of the DC/DC converter. This controller ensures that the inductor current, i_L, follows the reference,$i_{Lref}$, generated by the voltage controller. The closed-loop TF of the system, which is used to determine the PI controller gains, is represented in Equation (17):

$$i_L ={\displaystyle \frac{ \frac{1}{L_{DC}} \left(-V_{EV}\right)s}{s^2 + U\frac{K_{P i_L}}{L_{DC}}s + U \frac{K_{i i_L}}{L_{DC}}}}+ {\displaystyle \frac{ U\frac{K_{P i_L}}{L_{DC}} \left(s + \frac{K_{i i_L}}{K_{P i_L}}\right)}{s^2 + U\frac{K_{P i_L}}{L_{DC}}s + U \frac{K_{i i_L}}{L_{DC}}}}{ i}_{Lref}$$

(17)

By applying the limit as s tends to zero, it can be observed that the current,$i_L$ follows the reference current $i_{Lref}$, indicating that under steady-state conditions, the current follows the reference [43,44]. Comparing the denominator of the closed-loop TF (17) with the canonical form of a second-order system, the following equations for $K_{P i_L}$ and $K_{i i_L}$ are obtained:

$$\{\begin{array}{c}{K}_{P i_L} = {\displaystyle \frac{L_{DC}}{U}}2 {\xi }_{i_L}{{ \omega }_n}_{i_L}\\ K_{i i_L} ={\displaystyle \frac{L_{DC}}{U}} { \omega }_{n{i}_L}^2\end{array}$$

(18)

2.2.2. Battery Voltage Controller

When the battery voltage approaches the nominal voltage, the charging process ceases to be controlled by the charging power and is instead managed by the voltage compensator. This controller is designed to generate a current value, $i_{Lref}$, according to the required output voltage and the current determined by the charging power. Equation (15) was used to design the controller, expressing the relation between the variation in output voltage and the current $i_L$. The Laplace transform was applied, resulting in the following equation:

$$V_{EV}={\displaystyle \frac{1}{s{C}_{DC}}}\left(i_L-i_{EV}\right)$$

(19)

Using Equation (19), the control gains of the PI controller responsible for controlling the output voltage, $V_{EV}$, by forcing it to follow a reference given by the nominal voltage of the EV battery, can be computed based on a block diagram similar to Figure 4, where the closed-loop transfer function of a PI controller is given by Equation (20).

$$V_{EV}={\displaystyle \frac{ \frac{1}{C_{DC}} \left(-i_{EV}\right)s}{s^2+ \frac{K_{P V_{EV}}}{C_{DC}}s+ \frac{K_{i V_{EV}}}{C_{DC}}}}+{\displaystyle \frac{ \frac{K_{P V_{EV}}}{C_{DC}} \left(s+\frac{K_{i V_{EV}}}{K_{P V_{EV}}}\right)}{s^2+ \frac{K_{P V_{EV}}}{C_{DC}}s+ \frac{K_{i V_{EV}}}{C_{DC}}}}{ V}_{EVref}$$

(20)

To determine the control parameters of the PI controller, it is required to compare the transfer function (Equation (20)) with a second-order system, as shown in Equation (21) [43,44].

$$\{\begin{array}{c}{K}_{P V_{EV}}=C_{DC} 2 {\xi }_{V_{EV}} {{\omega }_n}_{V_{EV}}\\ K_{i V_{EV}}=C_{DC} { \omega }_{n{V}_{EV}}^2\end{array}$$

(21)

Here, ${{\omega }_n}_{V_{EV}}$ is defined as $2\pi$ divided by the system’s response time (${\displaystyle \frac{2\pi }{T_{V_{EV}}}}$), more precisely, the time it takes for the battery voltage, $V_{EV}$, to reach the reference value, ${ V}_{EVref}$, and the variable ${\xi }_{V_{EV}}$ represents the damping factor [44].

2.2.3. Frequency Droop Controller

The frequency droop control is based on the inverse relationship between isolated microgrid frequency and active power output. When the grid frequency is below the reference value, the generation increases the active power output, helping to raise the frequency. Conversely, when the frequency is above the reference, the generation reduces the active power output, contributing to frequency reduction [41].

This behavior mimics the natural response of synchronous generators and allows renewable sources and power electronic converters to participate in frequency regulation in a decentralized manner. The typical expression for frequency droop control is given by

$$P_{available}=(V_{DC}\ast i_{pv})(1+K_f\left(f_{ref}-f\right))$$

(22)

This expression defines how the system’s active power output $P_{available}$ to the EVs is adjusted based on the deviation between the grid frequency $f$ and its reference $f_{ref}$, with the droop coefficient $K_f$ controlling the responsiveness; larger values lead to stronger adjustments, while smaller values produce gentler corrections [41,46].

The frequency droop control mechanism regulates the active power exchanged with the electrical isolated microgrid based on deviations from the nominal frequency. When the isolated microgrid frequency falls below the reference value, the system injects current into the isolated microgrid to support frequency restoration, which consequently reduces $P_{available}$ to charge the EVs, as part of the current $i_{pv}$ is redirected to the isolated microgrid. Conversely, when the grid frequency exceeds the reference, the system absorbs current from the grid, resulting in an increase in the $P_{available}$. This dynamic adjustment contributes to grid frequency stability while managing the distribution of power between ancillary services and local loads [41,46].

This mechanism allows the system to contribute to the regulation of the isolated microgrid frequency, while dynamically adapting the charging power available to EVs.

3. Proposed Algorithm Frameworks

There are two algorithms proposed in this work: Power Sharing for massive EV charging and SEWP for charging a few EVs. The proposed Power Sharing algorithm is based on the traditional Power Sharing algorithm implemented on Wallbox chargers to manage the power from the grid, assigning the same amount of power to each EV [49]. On the other hand, the SEWP algorithm prioritizes the EV with the lowest SOC in comparison to the other EVs that are charging. Therefore, to better understand the proposed algorithms, in this section, we present a detailed flow diagram for each algorithm, explaining every step of the algorithm.

Besides using the algorithms, the PVCS provides the users with the possibility to charge their EVs exclusively with energy from the PV panels (denominated green mode), exclusively from the electrical grid (denominated red mode), or a mixture of both (denominated yellow mode).

3.1. Charging Algorithms

The proposed algorithms pursue three concurrent objectives: (i) maximize instantaneous utilization of PV power by allocating the net available PV energy to EVs charging in green mode, thereby minimizing curtailment; (ii) reduce state-of-charge (SOC) disparity across users by enforcing equal division in the Power Sharing policy and prioritizing the lowest-SOC EV in SEWP while distributing the residual power fairly; and (iii) strive to meet user-requested SOC targets within the declared parking duration.

3.1.1. Power Sharing Algorithm

A Power Sharing algorithm is a method for distributing power resources among different devices or units in a system, as shown in Figure 7. In essence, power sharing is a crucial mechanism for managing the complexities of modern power grids, ensuring stability, promoting equitable distribution, and optimizing resource utilization, particularly when dealing with grid energy limitations [49].

It can be applied to the management of the energy produced by PV panels to guarantee that every EV that is charging receives a certain amount of energy. The principle of the algorithm is to distribute the energy produced by the PV panels equally among all the charging EVs. More precisely, if an EV arrives at the parking lot and no other EVs are charging, then that EV receives all the power generated by the PV panels. If there is more than one EV charging in green mode, the power supplied to each charger is provided by dividing the power available for charging, $P_{available}$, by the number of EVs charging in green mode, $N_{green}$, as shown in Equation (23). It should be noted that the power supplied to the EV is always limited to the maximum power that the charger can provide.

$$P_{delivered}={\displaystyle \frac{P_{available}}{N_{green}}}$$

(23)

The flow diagram of Figure 8 illustrates the Power Sharing algorithm procedures for charging the EVs, highlighting the decision-making steps for assigning the power to all the EVs that are charging. Initially, the system evaluates the PV power and the number of EVs in the PVCS. Then the algorithm receives the SOC of the EV connected to the charger and the parameters introduced by the user, more precisely, the SOC the user wishes to achieve, and the EV charging time.

Finally, after the parameters are analyzed, the charger delivers the energy according to Equation (23). Therefore, the charger only ceases charging when the imposed charging time or the desired SOC (${SOC}_{ref}$) is reached.

3.1.2. SEWP Algorithm

The essence of the SEWP algorithm is to prioritize the EVs charging in green mode, $N_{green}$, according to the urgency of the charge, i.e., prioritize the EVs that wish to achieve a higher SOC in the shortest possible time. To apply this algorithm to manage the energy produced by the PV panels, it is imperative to adapt the priority. Instead of prioritizing the EVs with the ambition to achieve a higher SOC in the shortest possible time, the priority is given to EVs with the lowest SOC, which is updated after a certain period of time.

The EV with the highest priority receives a percentage ($Perc)$ of the $P_{available}$, indicated by $P_{with priority}$ on Equation (24). On the other hand, the EVs with lower priority receive the remaining portion of the power available,$\left(1-Perc\right)P_{available}$, which is equally shared among the non-priority vehicles, resulting in the expression $P_{without priority}$ on Equation (24). For the algorithm to be applicable, at least two EVs must be connected, so that one assumes the priority role while the others share the remaining power. Otherwise, when only a single EV is connected, the entire $P_{available}$ is allocated to its charging process.

$$\{\begin{array}{c}{P}_{with priority}=\left(Perc\right) P_{available}\\ P_{without priority}=\left({\displaystyle \frac{1-Perc}{N_{green}-1}}\right)P_{available}\end{array}$$

(24)

The flow diagram of Figure 9 illustrates the SEWP algorithm procedures for charging the EVs with the details explained above. Initially, the system evaluates the power available for charging the EVs and the number of EVs in the PVCS. Then the algorithm receives the SOC of the EV connected to the charger and the parameters introduced by the user, more precisely, the SOC the user wishes to achieve, and the EV charging time.

Finally, after the parameters are analyzed, the charger delivers the energy according to the SOC of the EVs, giving the $P_{with priority}$ to the EV with the lowest SOC and the $P_{without priority}$ to the other EVs. The charger only ceases charging when the imposed charging time or the ${SOC}_{ref}$ is reached.

3.2. Charging Modes

The system provides three charging modes: green, yellow, and red. The green mode allows the user to charge the EV exclusively with energy produced by the PV panels. On the other hand, the red mode only uses energy from the grid to the EV chargers. The yellow mode is a mix between green mode and red mode, i.e., the energy flows from the PV panels when there is energy available; otherwise, the energy flows from the electrical grid.

3.2.1. Green Mode (Solar Energy)

The green charging mode prioritizes the use of energy generated by photovoltaic PV panels, with the total available PV power distributed among the chargers operating under this mode. Under nominal operating conditions, specifically when the grid frequency remains stable, the power available for charging is typically supplied exclusively by the PV system. However, due to the implementation of frequency droop control, the active power exchanged with the grid may vary in response to frequency deviations, which can influence the allocation of PV energy and, consequently, the charging power delivered to the electric vehicles EVs.

To ensure efficient utilization of solar energy, energy management algorithms such as Power Sharing and SEWP are employed. These algorithms dynamically allocate the available PV power among the EV batteries, minimizing excess energy that would otherwise be injected into the grid.

3.2.2. Red Mode (Microgrid Energy)

Unlike the green charging mode, the red charging mode only allows the required power for charging to come from the electrical microgrid.

In this charging mode, the power applied to the charger is determined by the SOC that the user intends to achieve, the duration for which the EV remains connected to the charger, and the total battery capacity in kWh (calculated by multiplying the nominal battery voltage by the nominal battery capacity), as described by Equation (25).

$${\mathrm{P}}_{charger}=\left({\displaystyle \frac{{∆}_{SOC}}{100}}\right){\displaystyle \frac{E}{{\mathrm{t}}_{charging}}}$$

(25)

where ${∆}_{SOC}$ corresponds to the difference between the percentage of ${SOC}_{ref}$ that the user intends to achieve and the percentage of SOC the EV has, and ${\mathrm{t}}_{charging}$ is the time the vehicle will remain charging. $E$ represents the energy of the vehicle, calculated by multiplying the EV battery voltage by its capacity. It is important to note that the charger power ${\mathrm{P}}_{charger}$ is limited to the charger’s maximum power.

3.2.3. Yellow Mode (Mix of Solar with Microgrid Energy)

The yellow charging mode combines the operational principles of both the green and red modes, enabling EV charging through a hybrid energy supply that integrates PV generation with the electrical microgrid. In this configuration, the primary source of energy is the electrical microgrid, with the charging power limited to the level typically associated with residential charging [50].

PV energy is allocated to the EV only when there is power available from the PV panels ($P_{available}$ > 0) and no EVs are charging in green mode ($N_{green}$ = 0). If no PV power is available, or if any EVs are charging in green mode, the EVs in yellow mode continue to charge using energy supplied by the electrical microgrid, ensuring uninterrupted charging regardless of solar production. The energy supplied by the PV panels is managed according to the algorithm defined for the green charging mode, ensuring that PV generation is allocated following the same operational rules established for this mode.

4. Simulation Results

Several simulation cases are presented on MATLAB/Simulink R2023b to validate the effectiveness of the PVCS and its energy management algorithm. Two case studies are analyzed involving a PVCS with seven chargers and electric vehicles equipped with 32 kWh lithium-ion batteries with 400 V nominal voltage and a rated capacity of 80 Ah. The system includes 146 photovoltaic panels (each one with an area of 1.84 ${\mathrm{m}}^2$), totaling 51.8 kW of installed capacity. Although the system provides three charging modes, only the green mode was considered, as it corresponds to the scenario in which PV energy is the only source utilized and therefore requires algorithmic management.

Table 1. The parameters of the proposed system.

Parameter	Value
EV nominal voltage	400 V
EV rated capacity	80 Ah
EV internal resistance	0.05 Ω
Minimum SOC	20%
PV voltage at maximum power point, Vmp	37.8 V
PV current at maximum power point, Imp	9.39 A
$\mathrm{DC}/\mathrm{AC} \mathrm{converter} \mathrm{inductor} (L$)	1 mH
$\mathrm{DC}/\mathrm{AC} \mathrm{converter} \mathrm{capacitor} (C$)	10 mF
$\mathrm{AC}/\mathrm{DC} \mathrm{converter} \mathrm{inductor} (L_{charger}$)	3.57 mH
$\mathrm{DC}/\mathrm{DC} \mathrm{converter} \mathrm{inductor} (L_{DC}$)	3.73 mH
$\mathrm{DC}/\mathrm{DC} \mathrm{converter} \mathrm{capacitor} (C_{AC/DC}$)	10 mF
$\mathrm{DC}/\mathrm{DC} \mathrm{converter} \mathrm{capacitor} (C_{DC}$)	$1.25 \mathsf{\mu }$F
$\mathrm{Switching} \mathrm{frequency} (f_s$)	10 kHz
$K_{P i_{o1}}$	0.00889
$K_{i i_{o1}}$	1.97
$K_{P V_{DC}}$	−0.0885
$K_{i V_{DC}}$	−0.393
$K_V$	174.53
$K_{P i_L}$	0.0222
$K_{i i_L}$	49.0351
$K_{P V_{EV}}$	−0.0011
$K_{i V_{EV}}$	−0.0493
$K_f$	−0.005

presents the parameters of the system in use and the respective values.

To evaluate the algorithms’ performance, three irradiance scenarios are considered. The first simulates an ideal sunny day with unobstructed radiation, modeled by a smooth quadratic curve, as shown in Figure 10a. The second represents a partially cloudy environment, where irradiance fluctuates due to intermittent solar exposure, as indicated in Figure 10b. The third scenario models a heavily overcast day with extended periods of low or absent irradiance, as demonstrated in Figure 10c.

4.1. Effectiveness

The economic benefits of solar charging stations are evaluated from both the users’ and operators’ perspectives. For users, charging costs are lower than those of conventional public stations, offering substantial long-term savings. For operators, selling renewable energy at rates higher than those paid by the grid yields a more favorable return on investment.

To encourage the use of green mode, its price should match the cost of home charging. Yellow mode should be priced equivalent to slow public charging, approximately twice the home rate. Red mode should reflect fast public charging costs, about four times the home rate, to motivate users to prioritize green energy and minimize impact on the microgrid [50]. When vehicle inflow is low, surplus energy from photovoltaic panels can be redirected to the microgrid, with tariffs governed by market rates.

In the techno-economic analysis, the initial investment cost for implementing an electric vehicle charging system is EUR 38,544 for photovoltaic panels, EUR 17,655 for the DC/AC converter, EUR 50,316 for EV chargers, and EUR 37,200 for assembly and electrical installation, including protective equipment. This results in a total initial investment of EUR 143,715. The average annual energy production is estimated at 137,889 MWh, generating an annual revenue of EUR 27,577 with green charging at 0.2 EUR/kWh. The payback period is calculated to be approximately 6 years for a high utilization rate (90%), 7 years for a medium utilization rate (75%), and 11 years for a low utilization rate (50%). Including a stationary battery with a capacity of 188.89 kWh (at an estimated cost of EUR 138,188), the payback period doubles, bringing it closer to the lifespan of the photovoltaic panels and potentially exceeding the battery’s guaranteed lifespan.

In scenarios involving energy consumption from the electrical grid, as indicated by the phase shift between current $i_{o1}$ and voltage $v_{o1}$ in Figure 11a, there are no fluctuations in the current $i_{o1}$, ensuring a continuous and reliable supply of electricity to all users. During the transition period, at 1.5 s, when the EVs stop charging, and the energy begins to be injected into the grid, as shown in Figure 11b, the intelligent control system automatically adjusts energy distribution, allowing the transition to occur without any distortion in the current $i_{o1}$. When no EVs are present in the parking area, the surplus energy generated by the photovoltaic panels can be fed into the grid without causing current distortion, as illustrated in Figure 11c. This enables maximum utilization of solar energy while maintaining grid efficiency and stability, even during periods of low charging demand.

Figure 12 illustrates the condition in which the voltage $v_{o1}$ is elevated due to high PV generation, while the current $i_{o1}$ is phase-shifted relative to the voltage. This displacement reflects the exchange of reactive power associated with the droop control mechanism. Figure 12 highlights that, under these operating conditions, the DC/AC converter injects active power into the microgrid while simultaneously adjusting the reactive component to contribute to AC voltage regulation. The phase difference between voltage $v_{o1}$ and current $i_{o1}$ provides a clear visualization of the control action, demonstrating how the station supports grid stability without fully imposing the voltage reference.

4.2. Simulation of the Distributed Electric Vehicles Charging System

The simulation of the Distributed Electric Vehicles Charging System encompasses several key components, including the DC/AC converter, the DC/DC converter, and the overall behavior of the system in terms of AC current. Individual results will be presented for each component, highlighting their specific contributions to system performance. The simulation considers standard operating values, such as an AC voltage ($V_{oRMSref}$) of 230 V, a grid frequency reference ($f_{ref}$) of 50 Hz, and a battery voltage reference ($V_{EVref}$) of 400 V.

4.2.1. DC/AC Converter Simulation

To evaluate the proper behavior of the DC/AC converter and the MPPT algorithm, it is necessary to analyze the system’s response under two conditions: a constant variation in solar irradiance and a step change in irradiance. This approach allows for the assessment of whether the latter scenario introduces any distortion in the converted AC current, thereby validating the stability and accuracy of the energy conversion and tracking mechanisms under dynamic solar input conditions.

Figure 13 presents the variation in the current $i_{pv}$ under the scenario of Figure 10a, along with the corresponding AC current of phase 1 $i_{o1}$ at the output of the DC/AC converter. Additionally, it illustrates the system’s response to a step change in solar irradiance, showing both the resulting $i_{pv}$ current from the panels and the converted $i_{o1}$ current. This comparison enables the evaluation of the converter’s performance and the MPPT algorithm’s stability under both steady and abrupt changes in solar input.

Under ideal solar irradiance conditions, represented by a smooth parabolic curve, the PV panels generate the current $i_{pv}$ that mirrors this parabolic profile, Figure 13a. When converted to AC by the DC/AC converter, the waveform of $i_{o1}$ retains its proportionality to the irradiance but takes on a rounded, “candy-like” shape, Figure 13b, due to the modulation and filtering processes inherent to the conversion.

In a second scenario, a step change in irradiance was introduced at 3 s to simulate a sudden shift in solar input, as shown in Figure 13c. Although such abrupt variations are uncommon in real-world conditions, this test is used to evaluate the system’s robustness. The waveform stabilizes around 3.1 s, remaining free from oscillations shown in Figure 13d, which highlights the system’s ability to maintain high power quality even under extreme conditions. Throughout the variation in irradiance, the AC waveform remains distortion-free, with THD of 1% across all phases. This ensures compliance with grid standards and prevents fluctuations that could affect system performance.

This analysis of the converter’s response under sudden irradiance variations provides important insights into the system’s robustness and power quality. To complement this evaluation, Figure 14a presents the power curves for different solar radiation levels, highlighting the corresponding $V_{DC}$ operating points required for maximum power extraction from PV panels. Figure 14b shows the evolution of $V_{DC}$ under the scenario of Figure 10a, where the $V_{DC}$ dynamically follows the irradiance variation to ensure maximum power extraction.

The results confirm that the DC/AC converter responds rapidly and efficiently to changes in solar input, maintaining the PV panels at their optimal operating point. Consequently, the $V_{DC}$ remains proportional to the available irradiance, with a stable waveform and free from oscillations. These observations demonstrate the robustness of the MPPT algorithm and its effectiveness in preserving high power quality even under variable environmental conditions.

4.2.2. DC/DC Converter Simulation

The DC/DC converter is responsible for regulating the voltage levels between the voltage $U$ of the AC/DC converter and the battery storage system or the EVs being charged. This converter ensures that the voltage is appropriately adjusted to match the requirements of the connected devices. In this system, the voltage of the EV batteries is lower than the DC link voltage (output voltage of the AC/DC converter of the EV charger), and thus the converter reduces the voltage to match that of the battery. Figure 15 presents the results of the simulation of the buck-type DC/DC converter used in the charging process of EV batteries and shows the time evolution of the current $i_{EV}$ and the battery voltage $V_{EV}$, providing insight into the system’s behavior under controlled operating conditions. The analysis of that figure serves as a basis for assessing the performance of the implemented PI controller, as well as the system’s responsiveness to the typical demands of a smart charging profile.

The results of Figure 15 show that if energy is available from the PV panels, the charging current $i_{EV}$ remains constant during the initial phase, ensuring stable power delivery and maximizing solar utilization while the battery voltage $V_{EV}$ gradually increases. This behavior indicates that the system is operating in constant current mode, where the charging current $i_{EV}$ is kept fixed until the battery voltage $V_{EV}$ approaches its nominal value. This strategy is effective in ensuring a high charging rate without compromising battery integrity.

As the voltage $V_{EV}$ nears the nominal value, the PI output voltage controller activates, gradually reducing the charging current $i_{EV}$. The transition becomes evident in Figure 15 at approximately the 3 h mark, when the current $i_{EV}$ begins to decrease, and the battery voltage $V_{EV}$ stabilizes. This moment marks the initiation of a regulation phase, during which the control system actively limits current flow to prevent overvoltage conditions. By reducing electrical stress on internal battery components, this mechanism contributes significantly to battery health and extends its operational lifespan.

This charging profile, characterized by an initial constant current phase followed by a constant voltage phase, is typical of smart charging systems and aligns with best practices in lithium battery management. The results confirm that the DC/DC converter is capable of dynamically adjusting voltage and current levels, ensuring an efficient and safe charging process.

In EVs, battery degradation is influenced by factors such as temperature fluctuations, the SOC operating range, and the charging/discharging C-rate. Battery aging is primarily driven by calendar age, exposure to low or high temperatures, operation at high or low voltages, and high charge and discharge currents [51,52]. Using low C-rate chargers over longer charging periods can help preserve performance and extend battery lifetime. To further limit aging, charging should ideally occur in moderate-temperature environments and avoid sustaining very high SOC levels.

The efficiency of the DC/DC converter is further emphasized by its ability to minimize energy losses during the conversion process. This capability is particularly critical in scenarios where energy from photovoltaic panels is either stored or directly used for charging EVs. Figure 15 highlights the DC/DC converter’s effective voltage regulation, which significantly contributes to the overall efficiency of the PV charging system. By consistently regulating voltage, the DC/DC converter optimizes the EV charging process, ensuring a reliable and efficient power supply while minimizing energy losses. This is especially important for maximizing the utilization of energy generated by photovoltaic panels.

4.2.3. PVCS System Results

The simulation of the PVCS was carried out to analyze the behavior of different power distribution algorithms in terms of energy consumption from the electrical grid and the variation in the SOC when charging with each algorithm under various solar radiation scenarios. Two main algorithms were tested: Power Sharing and SEWP. The Power Sharing algorithm distributes the power generated by the photovoltaic panels proportionally among the EVs that are charging. The SEWP algorithm prioritizes EVs with lower SOC, distributing the energy more evenly among all EVs.

For the tests assessing the Power Sharing and SEWP algorithms, EVs are assumed to arrive at the parking lot early in the morning and to remain parked for an average of about two hours. This case represents a high EV-charging utilization level, since utilization is the key factor distinguishing the performance of the two power-allocation strategies. EVs are also assumed to request an SOC increase of roughly 20% to 30%. In addition, to evaluate differences in how the algorithms achieve the target final SOC, a 20% spread was imposed between the highest and lowest initial SOC among the EVs.

The choice of the priority percentage ($Perc)$ in the SEWP algorithm is a critical factor in the overall performance of the system and depends on the number of EVs charging. For example, when only two EVs are present, the $Perc$ for the EV with the lowest initial SOC should exceed 50%, thereby accelerating its charging process.

Following this, the determination of the $Perc$ was carried out by progressively increasing its value from 0%, in 1% increments, and analyzing the SOC evolution of the EVs under the condition that all chargers are occupied at the start of the charging process. Low values, such as 5% and 15%, fail to provide a meaningful improvement for the EV with the lowest SOC (EV 1) and instead lead to a significant imbalance in energy distribution. Under these settings, EV 1 receives substantially less energy compared to the remaining EVs, while non-priority EVs, such as EV 4 and EV 5, achieve considerably higher SOC levels. This disparity results in a pronounced gap between the SOC attained by EV 5 and that of EV 1, as illustrated in Figure 16a,b.

To reduce the SOC disparity between EV 5 and EV 1, the SEWP algorithm assigns a $Perc$ value of 20% of the total available energy to EV 1 when all chargers are occupied. This allocation represents an optimal compromise between prioritization and fairness, as illustrated in Figure 16c, allowing the most energy-deprived EV to benefit from preferential charging without imposing excessive penalties on the remaining EVs.

To provide a more comprehensive evaluation of each algorithm, Figure 17 presents a bar chart illustrating the evolution of the SOC of the EVs, with every EV charging simultaneously under the scenarios presented in Figure 10. In this chart, the blue bars represent the initial SOC of each EV, while the orange bars correspond to the SOC obtained under the solar irradiance scenario of Figure 10a, the gray bars indicate the SOC achieved in the scenario of Figure 10b, and the yellow bars depict the SOC reached in the scenario of Figure 10c.

The results obtained using the Power Sharing algorithm reveal a clear dependency between solar irradiance and the final SOC achieved by each EV. As the algorithm distributes the available photovoltaic PV power equally among all connected EVs, the SOC progression reflects both the initial charge level and the total energy available throughout the day.

For instance, EV 1, which began the charging period with a SOC of approximately 20%, reached a final SOC of 55% under the scenario of Figure 10a, 45% under the scenario of Figure 10b, and only 27% under the scenario of Figure 10c. Since Power Sharing does not prioritize vehicles with lower SOC, EV 1 remained at relatively low charge levels, which may reduce user satisfaction, particularly in scenarios with limited solar availability. In contrast, EV 5, which started with a higher SOC, achieved final SOCs of 75%, 65%, and 47% across the same scenarios. This outcome illustrates how vehicles with higher initial SOC benefit more from the equal distribution strategy, leading to higher satisfaction among these users.

In contrast, the SOC evolution observed for the SEWP algorithm reveals important insights into individual vehicle performance and user satisfaction across different irradiance scenarios. For instance, EV 1, which began the charging period with a SOC of approximately 20%, reached a final SOC of 70% under the scenario of Figure 10a, 55% under the scenario of Figure 10b, and 30% under the scenario of Figure 10c. This consistent prioritization of low-SOC vehicles ensured that EV 1 never remained critically undercharged, contributing to a high level of user satisfaction, particularly in scenarios with limited solar availability. In contrast, EV 5, which started with a relatively high SOC of 40%, reached final SOCs of 73%, 63%, and 46% across the same scenarios.

Quantitatively, both algorithms begin under the same conditions with the same initial SOC variance (36.57). However, their behavior diverges significantly by the end of the charging period. The Power Sharing algorithm maintains this variance throughout the entire process, indicating no improvement in SOC distribution among the electric vehicles. In contrast, the SEWP algorithm demonstrates a substantial reduction in SOC variance, achieving the value of 13.44 under the scenario of Figure 10a, 13.24 under the scenario of Figure 10b, and 26.50 under the scenario of Figure 10c. These results highlight the superior capability of the SEWP to enhance fairness and balance in energy allocation, particularly under favorable solar conditions.

Overall, the comparison highlights that while Power Sharing ensures equal distribution of PV energy, it tends to favor vehicles with higher initial SOC, leading to greater variation in user satisfaction. On the other hand, SEWP promotes a more equitable charging experience, where users of EVs with lower initial SOC perceive greater benefit without significantly compromising the charging progress of other vehicles. This balance contributes to higher overall satisfaction, particularly in shared or workplace charging environments where fairness and reliability are key operational goals.

To further illustrate these differences, Figure 18 presents the temporal evolution of SOC for the seven EVs of Figure 17. Unlike the bar chart, which summarizes the final SOC values, this representation allows the reader to observe the dynamics of charging throughout the day. By focusing on the scenario of Figure 10b, where PV generation is most constrained, the graph highlights how each algorithm manages scarce energy resources. This temporal perspective is particularly relevant for understanding the progression of EVs with lower initial SOC and for comparing the responsiveness of Power Sharing and SEWP in critical operating conditions.

The temporal variation reveals distinct charging behaviors between the two algorithms. Under the Power Sharing strategy (Figure 18a), the EV with lower initial SOC, EV 1, only reaches a SOC of 45% after 2 h of charging, evidencing the delayed response of this strategy. This uniform distribution of PV energy favors vehicles with higher initial SOC, such as EV 4 or EV 5, which consistently achieve higher charge levels. Consequently, user satisfaction varies significantly, with users of low-SOC EVs perceiving the system as less responsive to their needs.

In contrast, the SEWP algorithm (Figure 18b) demonstrates a prioritization of EVs with lower initial SOC. For instance, EV 1, which starts at a critically low level, exhibits a steeper charging curve and reaches a SOC of 45% after only 1 h and 45 min. This represents a 15 min improvement compared to the Power Sharing algorithm, under which the same SOC level is achieved after 2 h. Moreover, EV 1 continues charging under SEWP and reaches a SOC of 55% at the 2 h mark, further highlighting the efficiency of this strategy. Such behavior ensures that no EV remains undercharged for long periods, contributing to higher user satisfaction and reinforcing the perception of fairness in energy allocation. The homogeneity of SOC trajectories across the fleet further highlights SEWP’s effectiveness in balancing energy distribution, even under reduced irradiance conditions.

Overall, the comparison of temporal evolution curves provides deeper insights than final SOC values alone. It demonstrates that while Power Sharing ensures equal distribution of instantaneous PV power, SEWP enhances equity by prioritizing critical demand, resulting in more homogeneous charging outcomes and greater user confidence in shared infrastructure.

4.3. Impact of the PVCS on the Isolated Microgrid

To ensure stable operation of a solar-powered EV charging station connected to the electrical grid, the system employs voltage and frequency droop control strategies, whereby the station approximates the grid voltage reference by injecting or absorbing reactive power equivalent to about 30% of the apparent power; although it does not regulate the grid voltage in its entirety, this mechanism allows the station to contribute to voltage stabilization, and when multiple stations adopt the same approach, the grid voltage progressively converges toward the reference.

A simulation was conducted to evaluate the dynamic response of the voltage droop controller under grid voltage variations ranging from 207 V to 253 V, as illustrated in Figure 19a by the RMS voltage profile $v_{oRMS}$ with 230 V defined as the nominal reference. The droop control mechanism was activated only when microgrid voltage deviations exceeded ±5% and up to ±10% of the nominal value (230 V), ensuring operation within a predefined tolerance band. When the microgrid voltage dropped below this reference, the system injected reactive power into the microgrid to elevate the local voltage and support system stability, as shown in Figure 19b. Conversely, when the voltage exceeded the reference by 5% up to 8% of the nominal value, the system absorbed reactive power from the microgrid to mitigate overvoltage conditions. Beyond 8% of the nominal voltage, the reactive power absorption saturates at 30% of the apparent power, as shown in Figure 19b. The resulting reactive power profile exhibited a characteristic droop behavior, transitioning linearly from injection to absorption, thereby contributing to voltage regulation at the PCC.

To evaluate the voltage droop control strategy, a scenario was considered where the microgrid voltage at a microgrid point was monitored, as illustrated by the RMS voltage evolution at that point over time in Figure 20a. In Figure 20b, it is possible to observe the approximation of the microgrid voltage to its reference value; although it decreases by only 1 V, this represents a 4.3% reduction in the RMS voltage. Achieving full convergence would require multiple charging stations simultaneously contributing to voltage stabilization through reactive power support. Finally, Figure 20c clearly depicts the droop control mechanism, where reactive power is absorbed from the microgrid to reduce the voltage level, thereby demonstrating the system’s ability to maintain voltage stability.

In parallel, the frequency droop controller was assessed under grid frequency variations of $\pm 1$% of the nominal reference, between 49.5 Hz and 50.5 Hz, as shown in Figure 21b, with 50 Hz as the nominal reference. When the frequency fell below the nominal value, 50 Hz, the system injected active power into the isolated microgrid, which helped increase the frequency while decreasing the available power for EV charging. This additional power resulted from the combination of PV generation and a gain-proportional component linked to the injected current. Conversely, when the frequency exceeded 50 Hz, the system absorbed active power from the isolated microgrid to reduce the frequency, which led to an increase in the available charging power, as shown in Figure 21a.

Together, these droop control mechanisms enhance the operational flexibility and isolated microgrid-support capabilities of the charging station. By modulating the reactive power and charging current in response to voltage and frequency deviations, the system contributes to improved power quality, grid stability, and efficient integration of renewable energy. This approach reinforces the role of EV charging infrastructure as a distributed energy resource capable of actively participating in smart grid operations.

5. Conclusions

This study presented an innovative system for charging EVs using photovoltaic solar energy and energy management algorithms. Through the integration of DC/AC and AC/DC and DC/DC converters, as well as Power Sharing and SEWP algorithms, it was possible to maximize the use of solar energy and reduce dependence on electrical energy in isolated microgrids.

The DC/AC converter showed high efficiency and stable AC output with THD limited to 1%. The MPPT algorithm maintained optimal energy extraction under both gradual and abrupt irradiance variations. The DC/DC converter demonstrated effective voltage regulation and dynamic current control, operating in constant current mode followed by a constant voltage regulation. This behavior ensured stable power delivery, minimized energy losses, and protected EV battery integrity, supporting efficient solar utilization and extended battery lifespan.

Two main algorithms were developed: the Power Sharing algorithm and the SEWP algorithm. The comparison between the two algorithms highlights clear differences in charging outcomes and user satisfaction. Power Sharing, by equally distributing PV energy, favored vehicles with higher initial SOC, while leaving the lower SOC EV at relatively modest levels (27–55%), which reduced satisfaction under limited irradiance and maintained the variance across the charging process (36.57). In contrast, SEWP consistently prioritized low-SOC vehicles, enabling the EV with lower SOC to reach 30–70% across scenarios and reducing SOC dispersion across the fleet, and consequently achieving a small variance in the end of the charging process, 13.24–26.50. This approach ensured greater fairness and reliability, resulting in higher overall user satisfaction, particularly in environments where equitable energy distribution is essential.

The integration of voltage and frequency droop control strategies allowed the charging station to contribute to microgrid stability, with simulations confirming that the use of reactive power, limited to approximately 30% of the apparent power, effectively reduced the microgrid voltage by about 4.3% of the RMS voltage and adjusted the charging current in response to frequency deviations of $\pm 1$% of the nominal reference. This behavior enhances overall power quality and demonstrates the role of EV infrastructure as a supportive and dynamic element within isolated microgrid operations.

In summary, this study highlights the technical feasibility of using solar energy for EV charging, providing significant environmental benefits and contributing to the sustainability of the electrical system.